Simulation Analysis of Land-Use Spatial Conflict in a Geopark Based on the GMOP–Markov–PLUS Model: A Case Study of Yimengshan Geopark, China

Abstract

The foundation for accurately understanding regional land-use structures and pursuing the coordination of human–land relations is the scientific identification and simulation of temporal and spatial evolution patterns of land-use spatial conflict (LUSC). Based on the production–living–ecological space (PLES) perspective, a land-use spatial conflict identification and intensity diagnosis model (LUCSII) was constructed using a landscape ecology index. The methods of geographic information system (GIS), spatial autocorrelation analysis, and mathematical statistics were used to achieve the spatial pattern of LUSC over the last 20 years, and the GMOP–Markov–PLUS model was used to simulate the evolution of LUSC in the future under various scenarios. The results indicated that our established LUCSII could accurately identify potential land-use spatial conflict areas in geoparks. The GMOP–Markov–PLUS model constructed had also scientifically predicted the future land-use patterns under different scenarios, successfully demonstrating the changing process of spatial conflict pattern evolution. The research proposed three different plans for the long-term land use of YG, including ecological protection, economic development, and long-term development perspectives. Finally, the research further emphasized the importance for sustainable development of geoparks. More attention should be paid to the optimal allocation of land-use structure and the coordinated development of human–land relationships.

1. Introduction

The interplay and coupling of regional populations and the physical geography environment constitute a spatial conflict under the backdrop of territorial ecological alteration, and the land productivity, life, and ecological functions are becoming unbalanced [1]. Therefore, the 18th National Congress of the Communist Party of China clearly put forward the production–living–ecological space (PLES) development goal of “intensive and efficient production space, livable and moderate living space, and ecological space with beautiful mountains and clear waters”. At present, the problem of LUSC has attracted the attention of governments at all levels and all sectors of society. The quantitative identification of LUSC, its spatial patterns, and multi-scenario simulations have become hot issues in the field of geography and land science.

There has been a steady growth in research on geoparks worldwide since the early 1990s. The United Nations Educational, Scientific and Cultural Organization (UNESCO) defines geoparks as nationally protected areas with a number of geoheritage sites of particular importance, rarity, or aesthetic appeal [2]. Simultaneously, the International Union for the Conservation of Nature (IUCN) and the Sustainable Development Goals of the United Nations acknowledge the significance of safeguarding and promoting geoparks [3]. Additionally, the geoparks and tourism methods implemented by various countries for geological conservation and rural economic development, particularly in vulnerable environments, have been demonstrated to be effective [4,5]. At present, there is already some good research on geoparks, mainly focusing on the protection of geological heritage, geological education, geotourism, and geological diversity [3,6,7]. It is worth noting that sustainable management has been emphasized by some scholars as crucial for ensuring long-term benefits in rural development through geotourism [6,8]. Given the global support, geoparks require extensive scientific research to achieve sustainability [9]. Once the sustainable development of geoparks is widely recognized and valued, their protection within protected areas will become increasingly significant on local, national, and international natural and human welfare agendas [3].

In recent years, with the continuous increase in people’s demand for tourism, tourism products have gradually shifted from a single sightseeing category to a vacation category that integrates sightseeing tourism, vacation and leisure, health care and recuperation, and cultural entertainment [10]. Resorts, geoparks, and other specific areas organically integrate natural and cultural landscapes [11], becoming important places for professional upgrading of rural tourism. Geoparks are not only key protected areas for geological relic landscapes and ecological environments but also bases for conducting geological scientific research and popularization of education. The natural environment and socioeconomic issues caused by the development and protection of geological parks have received widespread attention. Land resources are an important carrier for the socioeconomic and ecological development of geological parks. Land-use types and structures within the region exhibit a diversified and composite trend driven by the planning and construction of supporting facilities such as transportation, catering, accommodation, shopping, and amusement facilities [12]. The land-use conflicts caused by these activities have become urgent problems to be solved for the sustainable development of geological parks.

Land-use conflict refers to inconsistency and disharmony among many stakeholders in the process of land resource use in terms of the methods, quantity, and other aspects of land use, as well as the contradictory state between various land-use methods and the environmental aspects [13]. Its connotation is the evolution of various interest conflicts and multiple land-use functions [14,15]. At present, the global academic community has achieved rich research results on land-use conflict, mainly focusing on the following four aspects. The first is research on the identification and classification of land-use conflict. There are three categories of methods for identifying land-use conflict: (1) qualitative identification methods such as participatory surveys, logical framework approach, and other qualitative identification methods [16,17]; (2) the comprehensive index method, which constructs a comprehensive index of land-use conflict through the pressure–state–response model and the complexity–vulnerability–stability model of land ecosystems [14,18,19]; this method is simple to operate and can reflect the relative strength of conflict; (3) the map factors overlay method, which is widely used for land-use conflict recognition and intensity diagnosis due to its accurate positioning [10,20]. In addition, the catastrophe series method, improved grey target model, minimum cumulative resistance model, and public participation GIS (PPGIS) have achieved good results in the identification of land-use conflict [21,22,23,24]. The classification system for land-use conflict varies due to differences in research perspectives, research purposes, and geographical location conditions [25,26]. The second is research on the causes and mechanisms of land-use conflict, in which the correlation studies reveal the impacts of land systems, land scarcity, land competition, and socioeconomic factors on land-use conflict [27,28,29], while pointing out natural conditions, population growth, historical basis, and economic development is an important factor affecting the evolution of land-use conflict [30,31,32,33]. In addition, global climate change and cross-regional flows of population and resources in the context of globalization have become new incentives for land-use conflict. The third is research on the mitigation and regulation of land-use conflict. Scholars use game theory, multi-objective planning, and genetic algorithms to regulate land-use conflict based on the perspective of land-use planning and management [34,35,36], focusing on alleviating multiple contradictions and on the decision-making equilibrium question, but it is difficult to achieve differentiated and targeted regulation of land-use conflict. With the development of remote sensing (RS), geographic information systems (GISs), and artificial intelligence technology, methods such as actor networks, participatory GISs, and spatial planning have become important methods for alleviating conflict [21,37,38]. Moreover, policies, systems, laws, and regulations are also important means of conflict regulation and control. The fourth is research on multi-scenario simulation and specific areas of land-use conflict, which are based on the future land-use simulation model (FLUS), the conversion of land use and its effects modeling framework simulation (CLUE-S), Markov cellular automata (CA), and other models [39,40,41]. The research on land-use conflict in specific areas such as mining areas, vacation areas, national park communities, and mine–grain mixed zones has received much attention [42,43,44]. Overall, research on land-use conflict is becoming increasingly comprehensive, greatly enriching the connotation and analysis methods of land-use conflict, providing a theoretical basis and scientific support for effectively alleviating land-use conflict and the contradictions between humans and land. The miniaturization of research scales has become a new research trend, and vacation areas and nature reserves with complex functional positioning have become new research hotspots. The evolution of future land-use conflicts is especially worthy of further exploration. With the popularization of the concept of economy–environment–society for sustainable development, exploring regional land zoning and use from the perspective of PLES has been accepted by many scholars [45]. Therefore, how to coordinate the functions of multiple land systems in these regions and optimize the territorial spatial pattern has become an important discussion topic to solve the conflict of land production, living space, and ecological space.

Delineating space for geological parks and allocating various spatial resources based on the perspective of the PLES has become an important lever for promoting ecological environmental protection, governance, and rural revitalization. Therefore, in order to conserve the ecological environment of the YG, the main aims of this study were as follows: (1) to construct an LUSC identification and intensity diagnosis model using the landscape pattern ecological index from the perspective of PLES, (2) to use the patch-generating land-use simulation (PLUS) model to construct different scenarios to simulate the evolution of future LUSC, and (3) to use GIS, spatial autocorrelation analysis, and other methods to reveal the spatial and temporal differentiation characteristics of LUSC from 2000 to 2030 in YG. This study can provide a reference for promoting the orderly use of land resources and building a harmonious relationship between humans and land in YG.

2. Materials and Methods

2.1. Study Area

The Yimengshan Geopark (YG) is in Linyi City, Shandong province, and ranges from 35°26′ N to 36°01′ N and from 117°43′ E to 118°14′ E (Figure 1). It involves seven counties—Mengyin, Feixian, Pingyi, Yishui, Yinan, Junan, and Linshu. There are hundreds of scenic spots including Mengshan Park, Diamond Park, Daigu Park, Menglianggu Park, and Yunmenghu Park. It is a comprehensive geological park with multiple geological heritage resources, with a total area of 1804.76 km². The YG is in a low mountain hillside area, surrounded by three mountains on the west, north, and south sides, with high terrain in the south and north and low terrain in the middle. The Yimeng Mountain region belongs to the continental monsoon climate of the East Asian warm temperate zone, with four distinct seasons. The average annual temperature ranges from 12 to 13.5 °C, and the average annual precipitation is 823.8 mm. The geological park is situated in an area rich in rivers and water resources. In 2022, the regional GDP was CNY 233.78 billion, the per capita GDP was CNY 42,528.65, and the population of permanent residents was 0.5497 million. With outstanding ecological service functions, this area is dominated by ecological space and serves as an important ecological barrier in southern Shandong province as well as an important water source protection area in the northern part of the Huai River Basin.

2.2. Data Sources

The data used in this study include land-use data, natural environment, location conditions, social economy, and other data (

Table 1. Data types and sources.

Data Type		Data Description	Data Sources
Land-use data	Land-use in 2000, 2010, and 2020	Raster, 30 m × 30 m	China’s land cover dynamic data sets from 1990 to 2021 (https://doi.org/10.5281/zenodo.5816591, accessed on 27 April 2023) (https://essd.copernicus.org/, accessed on 27 April 2023)
Natural environment data	Digital elevation model (DEM)		Geospatial Data Cloud (http://www.gscloud.cn, accessed on 27 April 2023)
	Slope
	Annual mean temperature	Raster, 1 km × 1 km	Resource and Environment Science and Data Center of the Chinese Academy of Sciences (http://www.resdc.cn/, accessed on 27 April 2023)
	Mean annual precipitation	Raster, 1 km × 1 km
	Soil type	Raster, 1 km × 1 km
	Soil texture	Raster, 1 km × 1 km
Location data	Distance from rivers	Vector	OpenStreetMap (https://www.openstreetmap.org/, accessed on 27 April 2023)
	Distance from main roads	Vector
	Distance from rural residential areas	Vector
	Distance from the county center	Vector
	Distance from hotels	Vector
	Distance from the government	Vector
Socioeconomic data	Yield of grain and grain price	—	Statistical yearbooks of county in the corresponding years in the Yimengshan Geopark and National Agricultural Statistics on Agricultural Products’ Cost and Income 2020
	Population density, gross domestic product (GDP), night light	Raster, 1 km × 1 km	Resource and Environment Science and Data Center of the Chinese Academy of Sciences (http://www.resdc.cn/, accessed on 27 April 2023)
Policy constraints	Open water surface	Raster, 30 m × 30 m

). The land-use data (2000, 2010, 2020) come from China’s land cover dynamic data set from 1990 to 2021 (https://doi.org/10.5281/zenodo.5816591, accessed on 27 April 2023) (https://essd.copernicus.org/, accessed on 27 April 2023) [46], and the resolution is 30 m by 30 m. Based on ArcGIS 10.4 software, land use is classified into six types: cultivated land, forest land, grassland, water body, construction land, and unused land. The data on grain prices and yields were sourced from the National Agricultural Statistics on Agricultural Products’ Cost and Income 2020 (https://data.cnki.net/, accessed on 27 April 2023), and the data of population and industrial gross domestic product come from the statistical yearbooks and statistical bulletin of the corresponding years of the counties and districts of Linyi city (http://tjj.linyi.gov.cn/, accessed on 27 April 2023). In addition, some missing indicator data were calculated from adjacent year data. Finally, the land-use data and the driving factors data were unified into WGS_1984_Albers projections with a spatial resolution of 30 m by 30 m, projected into a uniform frame of axes.

2.3. Research Methods

Based on the spatiotemporal evolution of LUSC in the YG, this study constructed a grey multi-objective optimization (GMOP)–Markov–PLUS coupling model to optimize land-use structures under four scenarios in 2030 and to calculate and analyze LUSC. The research framework is shown in Figure 2.

2.3.1. The Establishment of the PLES Spatial Classification System

Land has multifunctional properties. Based on previous research and actual situations [47], all land-use classification types were divided into production space, living space, and ecological space to establish a new land-use classification system to reflect the land-use spatiotemporal changes and spatial conflict status in YG (

Table 2. Land-use classification system of production–living–ecological space (PLES).

First Class	Secondary Class	Third Class	Fourth Class
Production space	Agricultural production space	Cultivated land (Gropland)	Paddy fields and dry fields
Living space	Industrial and mining production-living space	Construction land (Impervious)	Industrial, mining, transportation, and other construction land, urban land, rural residential land
	Urban living space
	Rural living space
Ecological space	Green ecological space	Forest land (forest, shrub), grassland (grassland and wetland), and unused land (barren)	Forest land, shrub land, sparse forest land, other forest land, high coverage grassland, medium coverage grassland, low coverage grassland, sandy land, saline alkali land, bare land, and bare rock gravel land
	Water ecological space	Water body (water)	Rivers, ditches, lakes, mudflat, beaches, marshes, reservoirs, and ponds

, Figure 3).

2.3.2. Construction of the Production–Living–Ecological Space Spatial Conflict Index Model

With rapid population growth, the impact of human activities on the natural environment has become increasingly perceptible, and land-use changes have led to changes in landscape structures [48]. The formation of landscape patterns reflects different ecological processes and affects the evolution of ecosystem landscapes. The landscape pattern index is a quantitative representation of the characteristics of land-use spatial patterns. In view of the complexity, vulnerability, stability, and other characteristics of the land-use ecosystem, combined with the resource and environmental characteristics in the study area, the landscape ecology index was constructed from the three aspects of complexity (P), vulnerability (V), and stability (S) to calculate the LUSC intensity of YG using the following formula:

$$LUCS=P+V-S$$

(1)

where LUCS is the comprehensive index of land-use spatial conflict, and P, V, and S are the spatial complexity index, spatial vulnerability index, and spatial stability index, respectively. In addition, based on ArcGIS 10.4 and Fragstats 4.2 software, the LUSC grades in YG were measured from 2000 to 2020. The equidistant method was used to classify the spatial conflict level into five categories: stable controllable [0, 0.2], mild conflict (0.2, 0.4], moderate conflict (0.4, 0.6], intense conflict (0.6, 0.8], and severe conflict (0.8, 1]. More details and guides for Fragstats software can be downloaded at http://www.umass.edu/landeco/research/fragstats/downloads/fragstats_downloads.html (accessed on 27 April 2023).

Spatial Complexity Index (LUP): The area’s weighted average fractal dimension (AWMPFD) is used in calculation. A higher LUP value indicates that the landscape unit has a greater chance of being disturbed by neighboring landscapes, and the land-use structure of the spatial unit is more complex.

$$LU{P}_{AWMPFD}={\displaystyle \sum _{i=1}^m{\displaystyle \sum _{j=1}^n\left[\frac{2\ln (0.25P_{ij})}{\ln (a_{ij})}\left(\frac{a_{ij}}{A}\right)\right]}}$$

(2)

Here, P_ij is the perimeter of the jth patch in the ith land-use category; a_ij is the area of the jth patch in the ith type of land use; A is the area of the spatial evaluation unit.

Spatial Vulnerability Index (LUV): The spatial vulnerability index reflects the vulnerability of the land-use system under external pressure interferences. Different landscape elements have different responses to external pressure. Based on the actual situation of YG, the spatial vulnerability indices of production space, living space, and ecological space were 1, 2, and 3, respectively.

$$LUV={\displaystyle \sum _{i=1}^n{F}_i}\times \frac{a_i}{S}$$

(3)

Here, n is the number of land types included in PLES land (n = 3); a_i is the area of each land type; F_i assigns a value for the vulnerability of different land types; and S is the area of the spatial evaluation unit.

Spatial Stability Index (LUS): the land-use stability index is calculated using the reciprocal of landscape fragmentation (PD), as follows:

$$LU{S}_{PD}=1-\frac{n_i}{A}$$

(4)

where n_i is the number of patches in land-use category i.

2.3.3. Spatial Autocorrelation Analysis

The spatial autocorrelation characteristics of LUSC in the YG were analyzed based on global and local analysis methods. In addition, the global Moran’s I index was used to discern the spatial clustering characteristics of LUSC in the YG, and local indicators of spatial association (LISA) was used to portray the spatial dependence and heterogeneity of spatial conflict intensity. The formulas were calculated as follows:

$$Moran’sI=\frac{n{\displaystyle \sum _{i=1}^n{\displaystyle \sum _{j=1}^n{W}_{ij}(x_i-\overline{x})(x_j-\overline{x})}}}{{\displaystyle \sum _{i=1}^n{\displaystyle \sum _{j=1}^n{W}_{ij}{\displaystyle \sum _{i=1}^n(x_i-\overline{x})(x_j-\overline{x}}{)}^2}}}$$

(5)

$$LIS{A}_i=\frac{\left(x_i-\overline{x}\right)}{{\displaystyle {\sum }_{i=1}^n\frac{{(x_i-\overline{x})}^2}{n}}}{\displaystyle {\sum }_{j=1}^n{W}_{ij}\left(x_j-\overline{x}\right)}$$

(6)

where LISA_i is the local spatial autocorrelation statistic of spatial unit i; n is the number of spatial units in the study area; x_i and x_j denote the LUSC intensity values of units i and j, respectively; $\overline{x}$ denotes the average value of LUSC intensity; and W_ij is the spatial weight matrix. The value of Moran’s I ranges from −1 to 1, with I > 0 indicating a positive spatial autocorrelation and I < 0 indicating a negative spatial autocorrelation. An LISA_i value greater than 0 means that the similar higher or lower values are clustered around the target grid, and the value of LISA_i being less than 0 indicates that a grid has high (low) value with low-value (high-value) neighbors.

2.3.4. Patch-Generating Land-Use Simulation

The patch-generating land-use simulation (PLUS) model compensates for the shortcomings of models such as CLUE-S and FLUS in mining transformation rules and simulating landscape dynamics. It can simulate multiple types of land-use changes at the patch level, and the simulation results can perfectly adapt to multi-objective optimization algorithms. This study uses the PLUS model to simulate the land-use patterns of YG in 2030 under different scenarios. The driving and limiting factors of land-use change refer to the research settings of Liang et al. [49] (Figure 4). Based on the land-use data of YG in 2010, the model accuracy was validated by fitting with the land-use data of 2020. Via a comparison with the actual land-use map, the overall accuracy and the Kappa coefficient were 89.88% and 0.888, respectively, which proved that the land-use simulation of the PLUS model can be effectively used for future land-use scenarios (Figure 5). More details and guides for PLUS software can be downloaded at https://github.com/HPSCIL/Patch-generating_Land_Use_Simulation_Model (accessed on 27 April 2023). The specific operation is as follows:

(a)

Simulation of the land-use spatial pattern

Based on the land-use data and socioeconomic natural data of YG in 2010 and 2020, this study explores land-use changes from 2010 to 2020, fits and optimizes land-use demand through Markov chain and GMOP models, and uses the PLUS model to simulate and predict the land-use distribution pattern of YG under the natural development scenario and the multi-objective optimization scenarios (Figure 5).

(b)

Design of multiple scenarios based on grey multi-objective optimization

The grey multi-objective optimization model can solve various uncertainties of objective functions and constraints in actual land use and provide decision makers with the best land-use optimization configuration plan [50]. To achieve ecological protection and coordinated economic development of the YG, four land-use scenarios were studied and designed: natural development, ecological priority, economic priority, and sustainable development scenarios [51]. By predicting the land-use demand under different scenarios, the simulation of the LUSC pattern for 2030 in YG can be achieved.

(1)

Natural development scenario (LED). The Markov chain predicts the total demand of land-use types in 2030 through the transfer probability between land-use types from 2010 to 2020.

$$P_{i,j}=\left[\begin{array}{ccc}{P}_{1,1}& \cdots & P_{1,n}\\ ⋮& \ddots & ⋮\\ P_{n,1}& \cdots & P_{n,n}\end{array}\right]$$

(7)

Here, n is the number of land-use types, P_ij is the probability of converting type i into type j, 0 ≤ P_i,j ≤ 1, and the sum of elements in each row is equal to 1.

(2)

Ecological priority scenario (EPD). The EPD ensures the priority development of the ecological benefits of the YG by setting the value of ecosystem service and maximizing the capacity of the ecological environment. Each land type can provide ecological benefits to the maximum extent. The ecosystem service value (ESV) and ecological capacity (EC) are ecological functions to measure ecological benefit. The calculation formulas for ecosystem service value (ESV) are as follows:

$$F_1(x)={\displaystyle \sum _{i=1}^{6}ES{V}_i}\times x_i$$

(8)

$$ES{V}_i={\displaystyle \sum _j^{9}a\times D}\times E_{ij}$$

(9)

where F₁(x) represents the total ecosystem service value provided by the land system; ESV_i is the ecosystem service value per unit area of land-use type i (10,000 CNY/hm²); a is the area of each cell network, which is 1 hm² in this study; D is the equivalence factors of ecosystem service value, and its calculation method is shown in reference [51]. Based on the national average grain price of CNY 2.19/kg from 2000 to 2020, D is calculated as CNY 1489.54/hm²; E_ij is the equivalent coefficient of the j-th ecological service value provided by the land-use type i, referring to the research settings of Xie et al. [52]. Finally, the total unit ESV values of cultivated land, forest land, grassland, water body, construction land, and unused land were calculated to be 0.58, 2.77, 2.32, 18.71, 0, and 0.03 (unit: CNY 10,000/hm²), respectively. If the construction land has no ecological benefits, the ESV of the construction land can be set to zero [53]. Therefore, the adjustment function of ESV can be expressed as the formula:

$$F_1(x)=0.58x_1+2.77x_2+2.32x_3+18.17x_4+0x_5+0.03x_6$$

(10)

The formula for calculating ecological carrying capacity is as follows:

$$F_2(x)=E{c}_i\times (100\%-12\%)$$

(11)

$$E{c}_i=Q_i\times Y_i$$

(12)

where F₂(x) represents the total ecological carrying capacity; Ec_i, Q_i, and Y_i are the ecological carrying capacity, equivalence factors, and yield factors of the land-use type i, respectively. Research by Wackernagel et al. [54] showed that retaining 12% of productive land is more conducive to biodiversity conservation. Referring to relevant research [54,55], we set the values of Q_i and Y_i, and calculated Ec_i. Formula (10) can then be rewritten as:

$$F_2(x)=5.35x_1+1.35x_2+0.20x_3+0.53x_4+5.35x_5+0x_6$$

(13)

In summary, the EPD multi-objective optimization function is expressed as max{F₁(x),F₂(x)}.

(3)

Economic priority scenario (RED). The goal of the RED is to maximize the economic benefits of various land-use types. Setting this scenario helps to scientifically grasp the potential risks of LUSC that YG may face, revealing the potential threat of economic priority development to the park’s development and protection, as well as its environmental carrying capacity. The formula is expressed as:

$$F_3(x)={\displaystyle \sum _{i=1}^{6}Ec{o}_i}\times x_i$$

(14)

where F₃(x) represents the total economic benefit; Eco_i is the economic benefit of land-use type i per unit area (CNY 10,000/hm²); x_i refers to the area (hm²) of land-use type i, with indices i ranging from 1 to 6 representing cultivated land, forest land, grassland, water body, construction land, and unused land. We estimate the economic benefits of cultivated land, forest land, grassland, and water body using the output values of agriculture, forestry, animal husbandry, and fisheries. The total output values of secondary and tertiary industries are an estimate of the economic benefits of construction land. The economic benefits of unused land are set to 0. Based on historical economic data from 2010 to 2020 and using the grey prediction model GM (1,1) to estimate Eco_i, Formula (14) can be rewritten as:

$$F_3(x)=3.60x_1+1.12x_2+18.04x_3+9.07x_4+89.34x_5+0x_6$$

(15)

In summary, the multi-objective optimization function in the RED scenario is expressed as max {F₃(x)}.

(4)

Sustainable development scenario (ESD). The ESD aims to build a harmonious human–land relationship in the YG by maximizing economic and ecological benefits. Its function is expressed as max{F₁(x),F₂(x),F₃(x)}.

The objective functions under EPD, RED, and ESD will be constrained by real-world conditions (

Table 3. Land-use structure constraints of the YG.

Constraint Factor	Constraints Condition/Unit	Formula Settings
Total factor	Total land area/hm2	$x_1+x_2+x_3+x_4+x_5+x_6=185372.28$
	Population size	The population density of cultivated land, forest land, and grassland is 3.01 people/hm2, the construction land is 8.32 people/hm2, and the population of the research area in 2030 is less than 717,936 people. $3.01\times \left(x_1+x_2+x_3\right)+8.32x_5\le 717936$
Macro planning	Agricultural land constraints/hm2	$x_1+x_2+x_3\le 163044$
	Construction land constraints/hm2	$17525.43\le x_5\le 24349.35$
	Cultivated land constraints/hm2	$120618.18\le x_1\le 123443.91$
	Forest land constraints/hm2	$0.46x_1+x_2+0.49x_3\ge 185372.28\times 50\%$
	Grassland constraints/hm2	$7791.77\le x_3\le 9694.08$
	Water body constraints/hm2	$4535.82\le x_4\le 4798.17$
Ecological needs	Biodiversity constraints/hm2	$\left(x_3+x_6\right)/185372.28\ge 3\%$
Non-negative constraint		$x_1\ge 0,x_2\ge 0,x_3\ge 0,x_4\ge 0,x_5\ge 0,x_6\ge 0$

). The constraint conditions for the multi-objective planning models are developed based on existing national spatial planning and related research. We imported the established objective function and constraint conditions into LINGO 18.0 modeling solution to obtain the land-use demand under EPD, RED, and ESD. Along with the land-use demand under LED, these are imported into the PLUS model to simulate the land-use patterns under four scenarios. Finally, we used the LUCSII model to simulate the LUSC pattern under four land-use scenarios of YG in 2030. More details and guides for LINGO 18.0 software can be downloaded at https://www.lindo.com/index.php/ls-downloads/try-lingo (accessed on 27 April 2023).

3. Results

3.1. Land-Use Spacial Conflict Spatial and Temporal Distribution Characteristics

The LUSCs were mainly stable controllable or had mild conflict. Marked changes in area occurred for different grades in the YG from 2000 to 2020 (

Table 4. Land-use spatial conflict changes in different grades in the Yimengshan Geopark (YG), 2000 and 2020.

Conflict Grade	2000		2020		2000–2020
	Area/km2	Proportion/%	Area/km2	Proportion/%	Area/km2	Rate of Change/%
Stable controllable	1037.38	55.96	817.37	44.09	−220.01	−21.21
Mild conflict	663.37	35.79	799.03	43.10	135.65	20.45
Moderate conflict	115.62	6.24	166.55	8.98	50.94	44.06
Intense conflict	27.74	1.50	44.13	2.38	16.39	59.08
Severe conflict	9.61	0.52	26.64	1.44	17.03	177.21

). During this period, the LUSC in the YG continued to intensify, manifesting as the contraction of the stable controllable area and the expansion of conflict zones in other grades. The stable controllable area decreased from 1037.38 km² in 2000 to 817.37 km² in 2020, with an average annual decrease of 1.06%; the area of mild conflict expanded by 135.65 km² or 20.45%. The area of moderate conflict increased by 50.94 km², with an average annual growth of 2.20%. In addition, the area of intense conflict increased by 59.08%, with an average annual growth of 2.95%, and the area of severe conflict has expanded by 1.77 times, with an average annual growth of 8.86%.

Obvious differences in the spatial distribution of the levels of LUSC were observed in the YG from 2000 to 2020, which demonstrate a general pattern of high in the middle and low in the exterior areas with a very high degree of coupling with the spatial distribution characteristics of construction land (Figure 6). The areas of severe conflict were mainly concentrated in the central urban area of Mengyin County and low and flat terrain areas such as southern Bailin Town. The areas of strong conflict were scattered around the construction land in each urban area, adjacent to the severe conflict areas, and along both sides of the river. The areas of mild and moderate conflict were distributed in the mountainous and hilly areas at the junction of Bailin Town, Liancheng Town, and Taoyin Town, as well as the ecotone between cultivated land and grass in the north.

According to the spatial conflict trends, urban and rural living spaces accounted for the majority of LUSC in the YG from 2000 and 2010. There was a large rural population during this time, and cultivated land and forestland served as the core production space elements of the rural regional system, providing the majority of farmers’ income. They were essential to the rivalry between various stakeholders, which caused a high degree of spatial conflict in rural areas. The demand for construction land has skyrocketed due to the region’s rapid urbanization, and some rural people have migrated into urban areas. The contradiction between humans and land in rural space was accompanied by population flow from rural areas to urban areas, making cities and towns a hotspot for spatial conflict. Spatial conflict hotspots started appearing in the water areas and the tourist attractions in the YG from 2010 to 2020. With the rapid development of the urban economy, the Linyi Municipal Government continues to plan and construct supporting facilities such as transportation, catering, lodging, shopping, and amusement facilities in the park, taking the successful application of the YG project as an opportunity. This has resulted in prominent effects on the natural environment and socioeconomic aspects of the YG and continual imbalances in ecological, production, and living spatial structures. As a result, the LUSC in tourist areas has been continually intensifying in the YG.

3.2. Spatial Agglomeration Characteristics of the Spatial Conflict

The Moran’s I index for 2000, 2010, and 2020 is 0.597, 0.610, and 0.624, respectively, and these values passed the significance test (Z > 1.96, p < 0.01). Thus, LUSC in the YG showed a significant positive spatial correlation, and spatial agglomeration is gradually strengthening.

To further explore the spatial clustering and dispersion of LUSC in the YG, we used local spatial autocorrelation to calculate the local Gi* index of land-use spatial conflict in the YG from 2000 to 2020. The corresponding spatial clustering map is shown in Figure 7. From 2000 to 2020, the LUSC agglomeration characteristics in the YG were high in the central south and low in the northeast, expanded outward along the road network. The high–high clusters shifted from the central region to the tourist attraction area of Duozhuang Town in the southwest, while the low–low clusters were found in the north central region, and the high–low and low–high clusters reduced. Figure 7 shows that in 2000, the high–high clusters were distributed in contiguous areas at the junction of Changlu Town, Gaodu Town, and Mengyin Street, as well as in the southern hilly areas. The low–low clusters were concentrated in Yedian Town, Daigu Town in the north, and the areas around Yunmeng Lake Wetland Park. Compared to 2010, the LUSC of YG in 2020 showed that the high–high clusters expanded and the low–low clusters contracted. The high–high agglomeration areas expanded in Mengyin County and Bailin Town, primarily because of the region’s frequent human economic construction activities, high proportion of living–production land, high spatial complexity of land use, and correspondingly prominent spatial fragility and low stability, which led to high spatial conflict and high value (high–high) clusters in these areas. The low–low clusters congregate around Duozhuang Town and Taoxu Town, which were affected by the terrain and have not undergone very marked changes in land-use types. Due to the minimal impact of human activity, the stability of the landscape remains high, and the low-value areas were dominantly impacted by the topography and distribution of water resources.

3.3. Multi-Scenario Simulations of Land-Use Spatial Conflict

The land-use structure and spatial conflict of YG under four scenarios in 2030 have similarities and differences (

Table 5. The land-use structure in the Yimengshan Geopark under different scenarios in 2030.

Scenario Type	Cultivated Land		Forest Land		Grassland		Water Body		Construction Land		Unused Land
	Area/km2	Proportion/%	Area/km2	Proportion/%	Area/km2	Proportion/%	Area/km2	Proportion/%	Area/km2	Proportion/%	Area/km2	Proportion/%
LED	1229.62	66.33	312.93	16.88	64.87	3.50	47.98	2.59	198.28	10.70	0.037	0.002
EPD	1234.44	66.59	299.07	16.13	95.03	5.13	48.32	2.61	176.83	9.54	0.027	0.001
RED	1206.18	65.07	364.18	19.65	55.66	3.00	48.40	2.61	179.25	9.67	0.045	0.002
ESD	1211.37	65.35	322.12	17.38	96.94	5.23	45.36	2.45	177.91	9.60	0.024	0.001

and

Table 6. The quantity situation of land-use spatial conflict in the Yimengshan Geopark under different scenarios in 2030.

Scenario Type	Stable Controllable		Mild Conflict		Moderate Conflict		Intense Conflict		Severe Conflict
	Area/km2	Proportion/%	Area/km2	Proportion/%	Area/km2	Proportion/%	Area/km2	Proportion/%	Area/km2	Proportion/%
LED	741.88	40.02	839.65	45.3	185.69	10.02	56.61	3.05	29.89	1.61
EPD	821.72	44.33	799.54	43.13	163.04	8.8	43.43	2.34	25.99	1.4
RED	700.35	37.78	894.32	48.24	178.69	9.64	51.17	2.76	29.19	1.57
ESD	767.26	41.39	840.86	45.36	174.26	9.4	42.80	2.31	28.54	1.54

). In 2030, the YG’s land-use structure will primarily consist of cultivated land, forestland, grasslands, and water bodies, with a small quantity of construction land and unused land. The LED construction land area is the largest among the four scenarios, at 198.28 km², or 10.70% of all land area. This scenario’s LUSC is considered more severe than the other three, with a severe conflict area of 29.89 km², or 1.61% of the total area. The RED has the least quantity of arable land, the largest quantity of forest land, and the second-highest percentage of construction land (9.67%) after the LED. The proportions of areas of intense and severe conflict in this scenario are 2.76% and 1.57%, respectively; the degree of spatial conflict is second only to the LED. The EPD has a construction land size of 176.83 km², accounting for 9.54% of the entire area, which is less than that of the other three scenarios. In this scenario, the fraction of areas of intense and severe conflict is 3.74% of the study area, and the degree of land-use spatial conflict is rather low. Among the four scenarios, the ESD features the most grassland and the least forest area, with building land accounting for 9.60% of the total area, second only to the EPD. In this scenario, the area of intense and severe conflict is 71.34 km², accounting for 3.85% of the total. Overall, under the LED and RED, YG focuses on economic development, with frequent human economic construction activities and noteworthy expansion of construction land through the conversion of cultivated land and grassland, which is concentrated in areas surrounding the county and towns. The EPD and ESD focus on environmental conservation and integrated socioeconomic development. The proportions of cultivated land and grassland with obvious ecological service functions are relatively high, and construction land expansion is slow. Moreover, the ESD has the minimum proportion of construction land and the smallest proportion of areas of intense and severe conflict, the totals of which are only 3.74% of the study area.

There are visible variations in the LUSC in the YG among the four 2030 scenarios (Figure 8), displaying overall patterns of high in the middle and low in the surrounding areas. The areas of severe conflict are mostly concentrated within county towns, important villages and towns, and major transit routes, covering areas such as Mengyin Street, Bailin Town, Changlu Town, and Duozhuang Town. Within the park, cultivated land, forest land, and grassland are broadly spread in low and flat terrain areas with little human activity and low levels of LUSC, which are mostly stable controllable zones. The intense and severe conflict regions have expanded toward Changlu Town, Bailin Town, and the county center in the LED and RED. The main expanding regions of intense and severe conflict include the border regions of Changlu Town, Liancheng Town, and Mengyin Street, as well as the rural neighboring regions of Duozhuang Town. The areas of intense and severe conflict in the EPD and ESD clearly contract toward areas such as Mengyin Street and Duozhuang Town. The main areas of intense and severe conflict are in the core of Mengyin County, the transportation arteries along Jiuzhai Township and Duozhuang Town, and in the urban–rural ecotone.

4. Discussion

4.1. Temporal and Spatial Pattern of Land-Use Spatial Conflict

Land is an essential element of human social development, and it supports the great majority of socioeconomic activity, among which multiple factors influence land-use conflict [42]. This study developed a measurement model for spatial conflicts based on complexity, vulnerability, and stability to uncover spatial disparities in park land-use spatial conflicts, as in earlier research on land-use conflicts [21,56]. According to the research, the increasing demand for land resources in densely inhabited and highly urbanized areas between 2000 and 2020 resulted in increased LUSC [43,57]. Using the GMOP–Markov–PLUS model to simulate future LUSC patterns, it was discovered that the land-use structure and spatial conflict of YG under four scenarios in 2030 have similarities and differences. The overall pattern of spatial conflict is high in the middle and low in the surrounding areas. The places of severe conflict are mostly concentrated in county towns, important villages or towns, and along major transit routes, covering Mengyin Street, Bailin Town, Changlu Town, and Duozhuang Town, among others. Within the park, cultivated land, forest land, and grassland are broadly spread in low and flat terrain areas with little human activity and low levels of LUSC, which are mostly stable controllable zones. Overall, ecological land has a low spatial conflict value; construction land, however, has a high spatial conflict value, which is consistent with previous research findings [58,59].

In comparison to prediction models such as the CLUE-S model and the FLUS model, the PLUS model can simulate the evolution of patch-level changes in various land-use types more effectively, and it has the advantage of revealing the contribution rates of land-use change’s driving factors, which allows it to simulate future land-use patterns with more scientific and logical outcomes [42,60]. As a result, the GMOP–Markov–PLUS model can better explain the relationship between the occurrence of LUSC and spatial changes under multiple future scenarios, providing a possibility for alleviating LUSC in the PLES of the park. Second, the study selected driving factors for simulating land-use structures from the natural environment, geographic location, and socioeconomic considerations, making the conclusions more objective and congruent with the findings of other scholars’ research [48,60]. The importance ranking of each driving factor for six types of land use was obtained by training the random forest (RF) algorithm in the PLUS model (Figure 9). The top four important factors driving the evolution of cultivated land, forest land, and construction land are elevation, slope, distance from rivers, and population distribution, which is consistent with the results of Li et al. and Guo et al. [56,61]. This implies that topographical characteristics, as well as population spatial distribution and flow, impact the patterns and evolution of spatial conflict, and humans play a prominent role in the LUSC process [40,62]. In addition, the GMOP–Markov–PLUS combination model not only can forecast LUSC under the natural development pattern, but it can also simulate spatial conflict in various circumstances. When the simulation results are compared, the optimal mode is chosen as the future development trend, which has major guiding significance.

4.2. Evaluation of the Multi-scenarios Simulation Effect of Land-Use Spatial Conflict

Trends of the LUSC grade changes in the YG (Figure 10) and of spatial intensity grades (Figure 11) are identified based on the LUSC patterns from 2000 to 2020 and the simulation results under four scenarios in 2030, to evaluate the effectiveness of LUSC resolution under each scenario.

From 2000 to 2020, the spatial conflict area of YG underwent remarkable regional expansion. The area of intense conflict increased from 27.74 km² in 2000 to 44.13 km² in 2020, and the area of severe conflict increased from 9.61 km² to 26.64 km², indicating that the human–land conflict in YG has intensified over the last 20 years. It is urgent to coordinate and balance the LUSC. Simulation results show that the LED continues the study area’s historical growth and change trajectory, causing unorganized expansion of the areas of severe and intense conflict. The simulation results of the RED scenario are slightly better than those of the LED, but the trend of intense and severe conflict zone expansion remains unchanged. In the EPD and ESD, the evolution curves of both intense and severe conflict areas show a downward trend, indicating that the land-use conflict situation has improved.

In summary, the EPD land-use strategy effectively reduces the degree of LUSC in the YG, whereas the ESD maintains the trend of land-use development in 2020. There is not a great difference in the spatial conflict zones of moderate and higher degree between these two scenarios. In the EPD, ecology is more important than economic development, and the most valuable asset in YG is the ecological environment. The natural environment in the area is magnificent, with excellent air quality, and its land-use landscape serves cultural, aesthetic, and other purposes. As a result, increasing the economy while focusing on environmental conservation is also a reasonable option. The ESD is a balanced development and protection relationship that not only protects the ecological environment and coordinates human–land relations but also promotes the economy. It is a scientific and rational model for development. In the next ten years, both the EPD and ESD will be viable options for easing and managing LUSC in the YG.

4.3. The Importance of Sustainable Development of Geoparks

More than 50% of global national parks, protected areas, and geological heritage is located on rural community land, which serves multiple functions such as ecological protection, agricultural production, housing, social and cultural attributes, and providing recreational services for the public [63,64]. At the same time, geological tourism has become an important source of income for many rural areas [7]. The YG has played a significant role in promoting local sustainable development and attracting a large number of tourists. According to the survey, the per capita disposable income of residents in the region has increased from CNY 12,355 in 2011 to CNY 28,808 in 2022, representing an increase of CNY 16,453 over the past 11 years and indicating a significant improvement in people’s living standards. Geoparks have stimulated socioeconomic activities and sustainable development by attracting an increasing number of visitors. By 2015, China’s geoparks received 438 million visitors, and tourism revenue reached CNY 149.279 billion [65]. Given the importance of geoparks for local geotourism and sustainable rural development, the further goal of geological parks is to maintain ecological stability and promote coordinated socioeconomic and environmental development.

Overall, the spatial conflict value between rural and urban areas in the Yimengshan Geopark is highest due to damage caused by service facilities, road traffic, and urban construction to the natural environment. This finding is consistent with previous research results [66,67]. Compared to other studies, this study offers fresh perspectives for the investigation of LUSC because it is based on PLES and establishes a conflict assessment methodology using the landscape pattern index [48,68]. Moreover, the Markov–GMOP–PLUS model with a dynamic perspective was proposed to simulate future land-use changes under natural social policy constraints. The model optimized the land-use structure in four different scenarios, taking into account the impact of various policies and socioeconomic development goals on future land factor allocation. This helps the government formulate more comprehensive territorial spatial plans to meet the needs of geoparks development and protection. In addition, the research findings have policy relevance and applicability and can serve as a guide for land-use planning in YG as well as other geological parks worldwide, despite some of the study’s limitations.

4.4. Limitations and Future Research

Based on the PLES perspective, this study constructed a LUSC identification and intensity diagnosis model using the landscape ecology index, enhancing and deepening the research on regional LUSC. This study relies on the practical requirements of development and environmental protection in the YG, which reveal the spatiotemporal differentiation characteristics of potential LUSC and use the GMOP–Markov–PLUS model to simulate the LUSC patterns under various future scenarios. Then, through analysis of the dynamic evolution of spatial conflicts in the YG, differential control strategies are set up through spatial conflict zoning. There has been limited study of land-use change and multiple scenario simulation in geological parks in LUSC research. Therefore, this study can fill gaps in the expression, type, intensity, and distribution characteristics of LUSC within the area.

In the spatial analysis method, the LUCSII model, which is based on the response of landscape patch characteristics to ecological environment protection, can quantitatively measure the spatial distribution characteristics of spatial conflicts and reveal the potential land-use risks in the geopark area [43,44]. The LUCSII model has obvious theoretical and practical usefulness, but it lacks the evidence of land-use adaptability and multifunction qualities [39,69]. The GMOP–Markov–PLUS combination model has some limitations due to the subjectivity of indicator determination and difficulty in quantifying these parameters, as well as limitations in the selection of socioeconomic and policy indicators within the geological park, which ignore competition and cooperation between townships [69]. As a result, the primary objective of this study is to simulate changes in LUSC using environmental, geographic, and socioeconomic factors. In the future, it will be possible to further quantify how policies affect the simulation and estimate the severity of land-use spatial conflicts based on the outcomes of land suitability assessments [42]. This not only reflects the characteristics of suitable land use but also the rivalries between different stakeholders for spatial resources.

5. Conclusions

(1)

From 2000 to 2020, the LUSC in the YG was mainly stable controllable or had mild conflict, with visible changes in the area at all grades. The central urban areas of Mengyin County and low-lying areas, such as southern Bailin Town, have seen the greatest increase in areas of severe conflict. The areas of intense conflict, which are second in scale, are distributed around construction land in various urban areas, adjacent to the severe conflict areas, and along both sides of the river.

(2)

LUSC in the YG showed an apparent positive spatial correlation, and spatial agglomeration was gradually strengthening. The high–high clusters were found in contiguous areas at the junction of Changlu Town, Gaodu Town, and Mengyin Street, as well as in the southern hilly areas. The low–low clusters were concentrated in Yedian Town, Daigu Town in the north, and areas surrounding Yunmeng Lake Wetland Park.

(3)

The land-use structure and spatial conflict of YG under four scenarios in 2030 exhibit similarities and differences. The expansion of construction land is obvious in the LED and RED, and the degree of spatial conflict is relatively high. In the next ten years, EPD and ESD will both be reasonable options for easing and controlling LUSC in YG.

A differentiated governance strategy of spatial conflict is proposed based on the manifestation, conflict degree, and spatiotemporal pattern characteristics of LUSC in YG, as well as the spatial pattern in different future scenarios. Areas with mild spatial conflict have a high degree of consistency between their tourism function positioning and land-use suitability, so land simply must be developed and used in accordance with the existing tourism function positioning. For areas with high levels of spatial conflict, strict protection of natural and cultural resources such as geological relics, mineral wonders, historical sites, cultural relics, and traditional villages should be strengthened, and the ecological function of tourist areas should be strengthened. The areas of severe conflict are found in urban and rural construction land, as well as tourist attraction facilities, which are the core functional regions of geological park tourism activities and residents’ everyday lives. The conflicts are particularly obvious in the disorderly expansion of construction land, as well as the spatial conflicts generated by the overlap between tourism of ecological land and tourism of construction land. In this portion of the region, it is necessary to determine the three zones (including urban space, agricultural space, and ecological space) and three lines (including urban development boundaries, permanent basic farmland, and ecological protection red lines) based on the demands of social and economic development, particularly the boundary red line for construction land growth (it is a limit for growth of construction land), and to plan production and living spaces. It is necessary to maintain regional ecological security and avoid future deterioration of the park’s ecological environment by performing well in terms of ecological isolation.